Mark Twain wrote:
I was concerned about the HD crashing.

Robert


The hard drive has "emergency retract". It's
supposed to use the spindle motor as a generator,
to provide enough power to move the arm back
to the resting position. Whether that's how it works,
who can be sure. The one and only web site (HGST) which
"explained" hard drive technology, is closed now.

The heads aren't really intended to "land" on the
platter, because of the possibility of stiction.
At one time, like my 250MB Quantum drive, there
was no ramp, and no place for the heads to go. And
the spindle would spin down, and the heads would
rest on a "textured" landing surface. And the heads
used to "stick", requiring the user to tap the
drive to try to free them up again. That's why
they put in a landing ramp.

On the old wash-tub (computer mainframe) drives,
they used violence at shutdown. There is a large
capacity for energy storage, and the capacitor
is dumped into the voice coil. The radial arm
retracts at high speed and hits a "stop" with
a resounding thud. It's a wonder the heads
don't fall off. Modern drives, when putting the
arm up the ramp, are gentle by comparison.

*******

I found a picture of a fan assembly similar to yours,
that shows screws for both levels. There are
some round-head screws holding the fan to the heatsink.

https://www.amazon.com/DW014-Heat-As...dp/B00ZSXMC7I/

If you're had one of the fan screws loose, you could
always examine it to see how it works. For better or
worse, the technique seems to be "metal screw wedges
in between aluminum surfaces", which could kick up
debris. It would be unnatural for them to use a tap
and thread the heatsink and use a machine screw instead.

For Dell, at least part of the design of their PC cases,
is for speed of assembly. But it could be that the
heatsink assembly is made in some other factory, in
which case driving screws in an interference fit into
the aluminum might be the way they do it. For the
companies that make these, what they want to do is
make a long bar of heatsinks with aluminum extrusion,
then saw off chunks of heatsink (like cutting a baloney).
They don't really want to do any more machining than that.
Cutting threads into the aluminum, would go against
their "philosophy".

Where it gets more interesting, is with copper core
aluminum heatsinks. Where a copper slug is inserted
into the aluminum. And that requires elevated temperature
or something. So that's a level above "baloney slicing"
on the factory floor. The copper slug method is one
level above plain aluminum heatsinks, and reduces the
thermal distance to the fins.

These are three ways to build CPU heatsinks. These
are in increasing order of cooling capacity.

1) plain aluminum
2) aluminum with copper slug in center (contacts CPU)
3) heatpipe cooler to distribute heat to the fins

And for the larger video cards, some of those use

4) vapor chamber (similar to heatpipe), a box with
a couple drops of alcohol or similar, near the base.

Heatpipes move heat better than a solid copper piece
of the same diameter. They work until you exceed the
"capacity" and the fluid inside can't condense because
the heat flux is too great for them. This would cause
the temperature to shoot up, because the transport
mechanism no longer works.

Video cards can use up to around 250W. Intel will be
coming out with a CPU soon, that uses more than that.
(It'll have at least two CPU dies inside, from a
regular processor.) It's a Xeon for servers or something.
It will be interesting to see how they cool it. Water
cooling used by hobbyists, wouldn't really be an option
for a server room. The cooler design must be reliable.

Paul